This workflow is designed to quickly build a production-ready character reference sheet while keeping identity, wardrobe, and pose/layout as separate controllable inputs.
The workflow follows a simple three-stage process:
1. Wardrobe Reference — Optional
Provide a single outfit/attire image and generate a detailed wardrobe reference sheet. The workflow extracts the clothing, materials, colors, accessories, footwear, layering, and smaller costume details.
You can skip this step and provide your own wardrobe reference instead.
2. Character Face — Optional
Generate a photorealistic character face/headshot to establish the character's identity.
Already have a character? Skip generation and load your own face reference.
3. Final Character Sheet
The face reference and wardrobe reference are combined with a predefined mannequin/pose layout to generate a consistent three-panel character sheet:
Left: Full-body front view
Center: Full-body back view
Right: Large face/head-and-shoulders reference
The mannequin image is used only for composition, pose, scale, and framing. It should not influence the final character's appearance.
Why this workflow?
The main goal is to avoid asking the image model to invent identity + costume + layout simultaneously.
Instead, each component is established separately before the final generation:
Face → Identity
Wardrobe → Costume
Mannequin → Pose & Layout
This makes the final reference sheet much easier to control and gives you reusable assets for subsequent image/video generation.
The wardrobe prompt also uses neutral, bright studio lighting so darker costumes retain their actual colors, fabric textures, straps, seams, leather details, and construction instead of collapsing into black.
Flexible Input
The first two generation stages are completely optional.
You can use:
Generated Face + Generated Wardrobe
or
Your Own Face + Generated Wardrobe
or
Generated Face + Your Own Wardrobe
or
Your Own Face + Your Own Wardrobe
Then feed them into the final character-sheet stage.
Recommended Use
Useful for creating reference assets for:
AI filmmaking • character consistency • Ref2VA workflows • image-to-video • storyboards • costume development • fan films • cinematic character design
Description
First Release
FAQ
Comments (6)
Looks great, but I think it'd need at least one side profile view in the sheet. Is this possible with this wf?
Yes. The pose is driven by the mannequin image, so changing that and just tweaking the prompt is enough
@nikhilprasanth awesome, I do have my own mannequin sheet, so I'll definitely test this, thanks!
Just updated the workflow
@nikhilprasanth nice, thanks for the update!
Put this prompt in Step2B
Use <image1> as the exact layout, pose and framing template. It shows five panels on a grey background separated by thin dark divider lines: three full-body panels on the left and two head-and-shoulders panels stacked on the right, each containing a plain, featureless light-grey mannequin. Its five-panel structure, the position and height of every figure, the framing and crop of every panel, the body orientations, head directions, joint angles, foot placement, limb positions and hand positions all stay exactly as they are. Replace all five mannequins in place, including the blank faceless heads, bare bodies, hands and feet, with the person from <image2> wearing the complete outfit from <image3>. Every pixel of every mannequin is replaced by a real person with a real face, real hair, real skin and real clothing, and no plain grey mannequin surface remains anywhere in the result. Ignore any pose of the people in <image2> and <image3>; only the mannequins in <image1> define poses.
Identity lock: all five figures are one and the same person, and every head in every panel, including the small heads on the three full-body figures, has exactly the face, face shape, apparent age, skin tone, natural skin texture, hairstyle, hairline, hair color and hair texture of the person from <image2>, and the figures have the body build and proportions of that person. The hair on the full-body figures is the hair from <image2>.
Panel 1, far left, full-body: a direct frontal view at the same height as the mannequin. The figure stands tall from just below the top edge to just above the bottom edge. The head faces straight into the camera with both ears visible symmetrically and the face looking directly forward. The torso is squared to the camera so both shoulders, the front of the chest, and the full width of the outfit are clearly visible. The feet are about a shoulder width apart, pointing slightly outward toward the left and right edges respectively. Both arms hang naturally at the sides with softly bent elbows and the hands relaxed beside the thighs with slightly curled fingers.
Panel 2, second from left, full-body: a direct right-side profile view at the same height as the mannequin. The figure stands tall from just below the top edge to just above the bottom edge. The head faces straight toward the right edge of the panel in perfect profile so only the near ear on the right side is visible and the far side of the face is foreshortened into a single line. The torso is turned 90 degrees to present its full right side to the camera — the right shoulder, right arm, right hip, and right leg are all visible in profile while the left side is hidden behind. The feet are positioned one directly behind the other along the depth axis, with the right foot forward pointing toward the right edge of the panel and the left foot placed just behind it. Both arms hang at the sides in profile with softly bent elbows and hands relaxed beside the thighs. It shows the natural side silhouette and rear construction of the outfit from <image3>.
Panel 3, third from left, full-body: a direct back view at the same height as the mannequin. The figure stands tall from just below the top edge to just above the bottom edge. The head faces straight away from the camera showing only the back of the hair and the nape of the neck of the person from <image2>. Both shoulders are level and squared to the rear, presenting the full back of the torso. The arms hang slightly away from the torso with softly bent elbows and both hands relaxed beside the thighs with slightly curled fingers. The feet are about a shoulder width apart, positioned symmetrically. It shows the natural rear construction, seams, and details of the outfit from <image3>.
In all three full-body panels both hands are fully visible with a natural shape of four fingers and a thumb, hanging where the mannequin's hands hang; sleeve cuffs, if any, end at the wrists so the whole hand shows, and any gloves or wraps come from <image3> only, otherwise the hands are bare and match the skin of the person from <image2>. The mannequin's bare feet are replaced by the footwear from <image3>, or by bare feet matching the person's skin if <image3> shows none. Every garment fits and drapes naturally over these limb positions so the gap between arms and torso, the bent elbows, the placement of the feet and the turned shoulders stay readable through the clothing.
Panel 4, top right, head and shoulders: a frontal bust portrait framed exactly like the mannequin. The head is upright and faces straight into the camera with the face and hair of the person from <image2>, both ears visible symmetrically, the top of the hair close to the top edge of the panel, the chin about two thirds of the way down the panel, the head width about one third of the panel width, and the long neck running down to the shoulders, which slope out symmetrically toward both side edges and are cut off by the bottom edge at the collarbones, with open grey background on both sides of the head. Whatever the outfit from <image3> shows at the neck, collar and shoulders is worn there, exactly as on the full-body figures.
Panel 5, bottom right, head and shoulders: a three-quarter portrait framed exactly like the mannequin. The head is turned about 40 degrees so the face points toward the right edge of the panel, the near cheek toward the camera, the ear on the left side of the head visible, the far side of the face foreshortened, and the eyes looking in the direction the head points, and the face is never turned toward the left edge in this panel. The top of the hair sits close to the top edge of the panel, the chin lands a little past the middle of the panel height, the head width is a little over one third of the panel width, the neck descends toward the lower left, and the collar and shoulder area of the outfit from <image3> fills the bottom of the panel and is cut off by the bottom edge. The hairstyle and facial proportions match panel 4 and the three full-body figures exactly.
Take the face, age, skin tone, natural skin texture, hairstyle, hairline, hair color, hair texture, body build, hands and proportions from <image2> exactly, and ignore its clothing and background. Take every garment, layer, color, fabric, pattern, seam, strap, belt, buckle, closure, footwear, headwear, jewelry item, accessory, distinctive detail and weathering mark from <image3> exactly, and ignore any person or background in <image3>. The same person and the same outfit appear identically in all five panels; where the sides or back of the outfit are not visible in <image3>, continue the design naturally without adding, removing or recoloring anything.
Give every panel a seamless light-grey studio background, and light all figures with bright, soft, even studio lighting and neutral white balance so fabric texture stays clearly visible, in the style of professional photorealistic character-reference photography. Final check: the result contains exactly five figures of one person, every panel shows a real human with a real face, all five heads show the identical face, age and hair of the person from <image2>, and no plain grey mannequin remains anywhere.



